titleBack to Pixels: Lifting Latent Video Diffusion into Pixel Space
authorsYanke Song, Shuo Chen, Chris Rockwell, Guande He, Nan Liu, Desai Xie, Tianpei Gu, Yong-Hyun Park, Aryan V S, Dennis Yuren Menn, Ming-Hsuan Yang, Xun Huang
date2026-11-12
venuearXiv (preprint, not peer reviewed)
idarXiv:2611.17336
affiliationRoblox Corporation
upvotes3 (Hugging Face Daily Papers, 2026-11-15)
linksabs pdf hf
cite
abstractFrontier video generation models work in a VAE latent space that compresses video in both space and time. This compression made training tractable, but now limits high-resolution, controllable, and interactive generation -- doubling resolution quadruples token count, conditions apply to groups of frames rather than individual ones, and each response must wait for the rest of its multi-frame block. We present PixVid, a simple yet effective method that lifts a pretrained latent video model to directly generate every frame in pixel space. PixVid keeps the diffusion transformer and replaces the VAE decoder with a convolutional head that renders its outputs into pixels. We find that predicting clean video rather than velocity, together with a larger noise shift, is key to quality. Finetuned from Wan2.1-14B, PixVid is competitive with the latent baseline within sim 1.3M training clips. Working framewise in pixels, it generates at 2times higher resolution with the same per-frame token count, keeps depth-conditioned video editing aligned under fast motion, and responds to each action at the next frame. Project page: https://chenshuo20.github.io/PixVid/

bibtex

@misc{song2026back,
  title={Back to Pixels: Lifting Latent Video Diffusion into Pixel Space},
  author={Yanke Song and Shuo Chen and Chris Rockwell and Guande He and Nan Liu and Desai Xie and Tianpei Gu and Yong-Hyun Park and Aryan V S and Dennis Yuren Menn and Ming-Hsuan Yang and Xun Huang},
  year={2026},
  eprint={2611.17336},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2611.17336}
}